The BBC is threatening legal action against AI search engine confusion, escalating tensions between publishers and technology companies to train artificial intelligence models using copyrighted news content.
According to Financial Times and other published reports, in a letter to Perplexity CEO Aravind Srinivas, the BBC alleges that Perplexity's “default AI model” will be trained using BBC materials, and that startups will stop all reductions in their content, remove copies used to develop AI, and propose claims for financial compensation.
According to the FT, this is the first time the BBC has called for a legal rep for AI companies to scrape content, reflecting concerns that freely available public sector content is being widely reused without permission. The station claims that some of the content is reproduced verbatim in confusion, and links to BBC articles have surfaced in search results. BBC executives claim that such practices undermined the BBC's reputation for impartial journalism and erode public trust, citing internal research that found 17% of confusion responses using BBC sources were important inaccurate or missing.
Perplexity is backed by Amazon founder Jeff Bezos and is reportedly finalizing the funding round at a $14 billion valuation. The FT reported that it dismissed the BBC's claim as “operational and opportunistic,” claiming that the broadcaster fundamentally misunderstands technology, the Internet and intellectual property laws. Rather than building or training basic models, the company uses its own system based on Meta's Llama to provide an interface for accessing Openai, Google and human models, claiming it is refined for accuracy. Perplexity has previously faced legal challenges from its News Corp subsidiaries, receiving halt and repeal letters from other major publishers, but has also signed revenue sharing agreements with outlets such as Time, Fortune and Der Spiegel.
“Therefore, it is very harmful to the BBC, and it undermines the reputation of the BBC's audience (including the UK fare payers who fund the BBC and undermine trust in the BBC.”
Recent reports by PYMNTS highlight the increased friction between generation AI companies and publishers over scraping of content that is frequently bewildered at the heart of such conflicts. PYMNTS reports on the launch of Perplexity's publisher's revenue sharing program in response to similar legal threats from the New York Times and Dow Jones, and the industry backlash.
